Mk. Demaray et Sn. Elliott, TEACHERS JUDGMENTS OF STUDENTS ACADEMIC FUNCTIONING - A COMPARISON OFACTUAL AND PREDICTED PERFORMANCES, School psychology quarterly, 13(1), 1998, pp. 8-24
The relationship between teachers' judgments of students' academic ach
ievement and students' performances on an achievement test was investi
gated. The major questions examined were: fa) How accurate are teacher
s' ratings of students' academic achievement? (b) What is the relation
ship between teachers' judgments, as operationalized by item predictio
ns on a standardized achievement test, and students' actual performanc
es on an achievement test? (c) Does a student's educational status, hi
gh-achieving versus low-achieving, affect the accuracy of teachers' ju
dgments? The study included 12 teachers who volunteered to participate
, and 47 students randomly selected from Wisconsin public schools. The
students consisted of 30 females and 17 males, ranging from first-to
fourth-grade levels. Teachers filled out the Academic Competence Scale
from the Social Skills Rating System-Teacher version (Gresham & Ellio
tt, 1990) and one questionnaire for each student. The questionnaire, d
eveloped by the authors of the: current study, required teachers to pr
edict how students would do on each item of the Kaufman Test of Educat
ional Achievement, Brief form (K-TEA; Kaufman & Kaufman, 1985). Studen
ts were then administered the K-TEA by a qualified examiner. Teachers'
judgments of students' academic achievement on the Academic Competenc
e Scale were correlated moderately high (r = .70) with students' actua
l K-TEA scores. Furthermore, mean percent agreement between teachers'
item predictions and students' actual performances on the K-TEA was mo
derately high (79%) (Kappa-coefficient of 67%). Lastly, there was part
ial support for the prediction that teachers were better predictors of
higher-achieving than lower-achieving students. Analyses demonstrated
a slightly higher percent agreement between teachers and students wit
h above-average K-TEA scores (81%) than between teachers and students
with below-average K-TEA scores (77%). This article concludes with a d
iscussion of the practical implications of these findings for psycholo
gists conducting academic assessments.